巻号一覧
53 巻 (2023)
- 4 号 p. 192-
- 3 号 p. 138-
- 2 号 p. 60-
- 1 号 p. 4-
53 巻, 1 号
選択された号の論文の4件中1～4を表示しています
- |<
- <
- 1
- >
- >|
招待論説
-
原田 拓, 西山 裕之, 秦野 亮原稿種別: 招待論説
2023 年 53 巻 1 号 p. 4-9
発行日: 2023/01/15
公開日: 2023/04/14
ジャーナル 認証ありThere is a growing need for machine learning technologies. In the past, the accuracy of the output results ( prediction results ) of machine learning models was the main focus of attention. Today, however, as machine learning technology is applied to various problem fields, it is required to explain not only the accuracy but also the basis on which the results were obtained. This is because machine learning models themselves are black boxes, and when data is input to a machine learning model, a certain output ( result ) is obtained, but it is difficult for users to understand why hat output ( result ) was obtained. For this background, a number of methods have been proposed to explain the basis of results obtained from machine learning models, and they are called “ eXplainable Artificial Intelligence：XAI ”. A variety of research is being conducted, from basic research to applied research. This paper introduces some of the explanation methods of XAI.抄録全体を表示PDF形式でダウンロード (1542K)
研究室紹介
-
佐野 夏樹原稿種別: 研究室紹介
2023 年 53 巻 1 号 p. 10-13
発行日: 2023/01/15
公開日: 2023/04/14
ジャーナル 認証ありIn this article, we introduce the profile of the faculty of informatics, Tokyo University of Information Sciences （TUIS） and an education program in TUIS. And, we introduce an example of activity of SANO laboratory. SANO laboratory developed an education software of the Monty Hall problem for exhibition at school festival of TUIS and gathered about eighty people into our site and had them experience the software program. The developed software program can help intuitively understand the Monty Hall problem through dialogue interface and Monte Carlo simulation. In this article, we briefly summarize Monty Hall problem and the developed education software.
In addition, we introduce results of study concerning generation method of synthetic data by principal component analysis. We generated synthetic data from the anonymized data from a national survey of family income and expenditure in Japan by the proposed method and evaluated the utility and risk of the generated data.抄録全体を表示PDF形式でダウンロード (1195K)
講演概要
-
～なぜ職場で問題解決が進まないのか～古谷 健夫原稿種別: 講演概要 第130回クオリティトーク
2023 年 53 巻 1 号 p. 14-19
発行日: 2023/01/15
公開日: 2023/04/14
ジャーナル 認証ありFor many years, in addition to supporting various KAIZEN activities in companies ( manufacturing and service industries ), I have also provided practical support for problem solving for hospital doctors, prefectural government officials, NPO representatives, etc. In various situations, I found that most people did not have a basic understanding of problem solving. In addition, I found that there are common factors that prevent problem solving from proceeding well in all workplaces. In this paper, I clarified the factors that prevent problem solving from proceeding well, and summarized what I would like people who are working on problem solving to pay special attention to. I hope that problem solving will spread and permeate society and increase the international competitiveness of Japan.抄録全体を表示PDF形式でダウンロード (1823K)
ルポルタージュ
-
佃 康司原稿種別: ルポルタージュ
2023 年 53 巻 1 号 p. 20-27
発行日: 2023/01/15
公開日: 2023/04/14
ジャーナル 認証ありPDF形式でダウンロード (864K)
- |<
- <
- 1
- >
- >|